Monitoring surface deformation of the Harmalière and Avignonet landslides (France) using remote sensing and climate data

Abstract

Landslides pose significant risks to human settlements and infrastructures. Slow-moving landslides, while less immediately threatening, can develop into rapid and destructive events, making their study essential for understanding their geomorphological processes. This research focuses on the Harmalière and Avignonet landslide in the French Alps, two adjacent slow-moving landslides with distinct behaviours. The Harmalière landslide has experienced periods of rapid advancement and morphological changes due to multiple large-scale reactivations, while the Avignonet landslide has remained relatively stable with slow retrogression rates. Although quite some research has already been done on these landslides, the usage of remote sensing data is limited. To analyse these landslides, a multi-disciplinary approach was applied, combining Unmanned Aerial Vehicle (UAV) imagery, Light Detection and Ranging (LiDAR) data, historical aerial images, and climate data. This was ultimately done to investigate landslide dynamics, morphological changes, and climate triggers. This study showed that the Harmalière landslide has experienced several distinct reactivation events which caused a headscarp retreat of 130 metres, with sediment displacements exceeding 106 m3 during major reactivations. In contrast, the Avignonet landslide showed slower rates of deformation and less morphological evolution. The UAV-derived DEMs provided high-resolution insights into recent landscape evolution and the LiDAR DEMs highlighted the influence of vegetation on monitoring. The historical images enabled the reconstruction of past reactivation phases of the Harmalière landslide and their associated volumes. In total nine different reactivation events were found in the literature, five of which were analysed in detail. The events of March 1981 and June 2016 were especially large (~ 2 x 106 m3 ). The landscape is shaped by geomorphological processes including intense erosion at the headscarp and substantial deposition in the accumulation zone. Climate analysis identified rapid snowmelt and intense rainfall as the primary triggers for the reactivation events. This research highlights the value of combining UAV, LiDAR, and historical imagery with climate analysis to improve our knowledge of landslide dynamics. All these different sources have shown to be an effective way to study a landslide. The UAV and LiDAR images show significantly higher quality compared to the historical images. However, these historical photographs remain valuable for long-term reconstruction.